27 / 76
CARDIOVASCULAR JOURNAL OF AFRICA • Volume 29, No 3, May/June 2018
AFRICA
157
of observed degree of atherosclerosis among HIV-infected
participants. Briefly, mean cIMTwas ascertained at the distal 1 cm
of the common carotid arteries on images obtained in B-mode
along the lateral, anterior and posterior longitudinal sections of
the common carotid artery bilaterally (right and left common
carotid) as per the 2008 American Society of Echocardiography
Carotid Intima–Media Thickness Task Force protocol.
27
A
Sonosite M-turbo
©
ultrasound machine (FUJIFILM Sonosite
Inc, Bothell, WA, USA) connected to an 8–12-MHz linear probe
was used to obtain still images at the beginning of the R wave,
and sonocal
©
(version 5 of 2011) was used to measure cIMT in
auto-mode as per the manufacturer’s instructions.
Statistical analysis
Baseline characteristics were compared using the two-samples
t
-test for continuous variables and Fisher’s exact test for
categorical variables. All biomarkers and cIMT were initially
compared between HIV-infected versus HIV-uninfected
participants. The association between HIV status and level
of each biomarker of endothelial dysfunction was assessed in
univariate and multivariate linear regression, with the following
variables selected
a priori
in the multivariate model: age, gender,
BMI, mean arterial pressure (MAP), total cholesterol and
glycosylated haemoglobin levels, smoking status (current
non-smoker vs smoker) and statin use. Both the biomarkers
of endothelial dysfunction and sCD163 and IL-6 levels were
assessed for deviations from the normal distribution and if noted
to be skewed, data were log transformed.
During initial model building to assess the association
between HIV status and endothelial dysfunction, observed
associations were assessed both visually in graphical plot and
formally (the latter by including a quadratic term in the model
and testing for significance). These additional assessments are
not reported for ICAM-1 and VCAM-1 because the overall level
of significance of the association was unchanged, and similarly it
was not reported for E-selectin due to high rates of missing data.
Finally, among HIV-infected participants only, IL-6 and
sCD163 were assessed for associations with endothelial
dysfunction in univariate models, and then in multivariate
models adjusted for each other (IL-6 vs sCD163) and cIMT
(as a composite surrogate marker of observed burden of
atherosclerosis). If any outliers were noted during preliminary
data review (which occurred for IL-6 and log sCD163), sensitivity
analyses were performed excluding these observations, and
findings were reported if they were different. Furthermore,
the three covariates (IL-6, sCD163 and cIMT) were assessed
for collinearity and any detected high correlation (
>
0.3) was
considered to not impact on the analyses if model estimates
remained similar between univariate and multivariate models
(i.e. estimates were in the same direction and with similar
significance level).
As in the initial model building above, the association
between predictors (IL-6, sCD163 and cIMT) for individual
models (ICAM-1, VCAM-1, E-selectin) was assessed both
visually in graphical plot and formally (the latter by including
a quadratic term in the model and testing for significance). A
p
-value
<
0.05 was considered significant in all analyses. All
analysis was done in SAS
©
, version 9.4 (SAS Institute, Cary,
North Carolina, USA).
Results
Among the participants, 112 were HIV infected with HIV-1
RNA
<
400 copies/ml (51% were female, mean age 40
±
5 years)
and 84 were HIV-uninfected controls (45% female, mean age 38
±
5 years). HIV-infected participants had a mean HIV disease
duration of 9.8
±
3.1 years and had been on ART for 8.5
±
2.7
years, with the majority (73%) still on first-line NNRTI-based
ART as per the Botswana national ART guidelines. In addition
Table 1. Demographics and clinical characteristics of HIV-infected
participants and HIV-uninfected controls
Variables
HIV-
infected,
n
=
112
a
HIV-
uninfected,
n
=
84
a
p
-value
Age (years)
40
±
5
38
±
5
0.02
Female,
n
(%)
57 (51)
38 (45)
0.47
Cigarette smoking,
n
(%)
Never
70 (63)
56 (67)
<
0.01
Prior smoker
36 (32)
12 (14)
Current smoker
6 (5)
16 (19)
Family history
Myocardial infarction,
n
(%)
1 (1)
4 (5)
0.17
Stroke,
n
(%)
8 (7)
18 (21)
0.01
Personal history
Diabetes mellitus,
n
(%)
1 (1)
0 (0)
1.00
Hypertension,
n
(%)
15 (13)
10 (12)
0.83
Chronic kidney disease,
n
(%)
4 (3)
0 (0)
0.14
Dyslipidaemia,
n
(%)
9 (8)
2 (2)
0.12
Medications (current)
Anti-hypertensives,
n
(%)
15 (13)
9 (11)
0.66
Statins,
n
(%)
6 (5)
1 (1)
0.24
Anthropometric data
Systolic blood pressure (mmHg)
129.7
±
15.1 130.4
±
17.0 0.75
Diastolic blood pressure (mmHg)
84.6
±
11.5 84.6
±
13.3 0.97
Body mass index (kg/m
2
)
Underweight (
<
18.5),
n
(%)
10 (9)
8 (10)
0.98
Normal weight (18.5–
<
25),
n
(%)
58 (52)
43 (51)
Overweight (25–
<
30),
n
(%)
28 (25)
23 (27)
Obese (≥ 30),
n
(%)
15 (14)
10 (12)
CVD risk laboratory tests
Total cholesterol (mmol/l)
4.8
±
1.2
4.5
±
1.2
0.06
LDL cholesterol (mmol/l)
2.9
±
1.0
2.6
±
1.0
0.02
HDL cholesterol (mmol/l)
1.4
±
0.5
1.4
±
0.4
0.87
Triglycerides (mmol/l)
1.4
±
0.8
1.2
±
1.5
0.41
Glycosylated haemoglobin (g/dl)
5.3
±
0.4
5.5
±
1.0
0.01
HIV parameters
HIV disease duration (years)
9.8
±
3.1
Duration on ART (years)
8.5
±
2.7
Nadir CD4 count (cells/μl)
113
±
74
Baseline CD4 count
c
(cells/μl)
117
±
72
Current CD4 count (cells/μl)
553
±
251
Proportion with undetectable VL,
n
(%)
100
Time since most recent VL (months)
3.2
±
2.0
NNRTI-based ART,
n
(%)
82 (73)
PI-based ART,
n
(%)
29 (26)
HIV, human immunodeficiency virus; ACE, angiotensin converting enzyme;
HMG-Co A, 3-hydroxy-3-methylglutaryl-coenzyme A; CVD, cardiovascular
disease; LDL, low-density lipoprotein; HDL, high-density lipoprotein; baseline
CD4 count, pre-ART CD4 count; VL, viral load; NNRTI, non-nucleoside
reverse transcriptase inhibitor; PI, protease inhibitor
a
Means (standard deviations) for continuous measures; counts (percentages) for
categorical measures.
b
Continuous variables were compared using the two-samples
t
-test and categori-
cal variables were compared using Fischer’s exact test.
c
Baseline CD4 count indicates pre-ART CD4 count.